storm@cs.mcgill.ca (Marc WANDSCHNEIDER) (04/15/91)
I can't decide whether to get a large hard drive this summer or a medium sized hard drive and a math chip. Which of course is going to lead to the question, do I really need a math chip? I will not be using AutoCAD, and such, but will be using the windows 3.0 SDK which apparantly needs the chip, and I will be doing much other intense comp- iling efforts. However, apart from that and graphics such as Fractals, do you think it's worth the $350 I'm going to have to for over for it...? Opinions (reasonable ones ie no "get a mac") would be appreciated. ./*- -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ storm@cs.mcgill.ca McGill University It's 11pm, do YOU Marc Wandschneider Montreal, CANADA know what time it is? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
jcburt@ipsun.larc.nasa.gov (John Burton) (04/17/91)
In article <1991Apr15.045028.7247@cs.mcgill.ca> storm@cs.mcgill.ca (Marc WANDSCHNEIDER) writes: > >I can't decide whether to get a large hard drive this summer or a medium sized >hard drive and a math chip. Which of course is going to lead to the question, >do I really need a math chip? > >I will not be using AutoCAD, and such, but will be using the windows 3.0 SDK >which apparantly needs the chip, and I will be doing much other intense comp- >iling efforts. > >However, apart from that and graphics such as Fractals, do you think it's worth >the $350 I'm going to have to for over for it...? > >Opinions (reasonable ones ie no "get a mac") would be appreciated. > >./*- >-- >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >storm@cs.mcgill.ca McGill University It's 11pm, do YOU >Marc Wandschneider Montreal, CANADA know what time it is? >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Compiling a program generally does not make extensive use of a math coprocessor. Fractals and more advanced graphics can make extensive use of a math coprocessor. Graphics primitives (line/circle drawing, area filling, etc) can be written to make use of INTEGER ONLY math (in which case a math chip does NO good, since it only handles floating point operations). If you're using a package of graphics routines (such as those supplied with TURBO C), well, I'm not sure, it depends on how the routines are coded (sometimes these packages have different flavors of the routines depending on if you have a math chip or not). The answer? flip a coin...it really depends on how tight the disk space problem is...if you're really strapped for disk space, spend a little extra on the disk (for $350 you can get ~60 MB of disk). If the extra disk space is not required but just a "nicety", get the co-processor. BTW: Just glancing at price trends for disk space & coprocessors, the price of disk storage ($/MB) has been dropping *much* faster than the price of coprocessors, so it might be a good idea to go ahead and get the coprocessor, but wait 6 months for the disk (if you *can* wait for the storage). Or you could just wait 6 months for both & see what (if any) effect the introduction of the 80586 and the AMD/Intel court fight will have on the price of coprocessors... Hope this helps... John +--------------------------------------------------------------------+ | John Burton | | G & A Technical Software | | jcburt@gatsibm.larc.nasa.gov | | jcburt@cs.wm.edu | | | | Disclaimer: Hey, what can I say...These are *my* views, not those | | of anyone else, be they employer, school, or government| +--------------------------------------------------------------------+